Sechenov University specialists have created a cloud-based medical data platform to study new treatment methods using AI

2023: Building a Cloud-Based Healthcare Data Platform

Specialists from Sechenov University, Yandex Cloud and Beltel Datanomics have completed the creation of a medical data platform. 18 million medical documents have been uploaded to the platform. The solution will help researchers, doctors, students and graduate students of the Sechenov University of the Ministry of Health of the Russian Federation conduct scientific research, evaluate the effectiveness of treatment methods in different groups of patients, and IT teams - create artificial intelligence services in the field of health. More than 1000 specialists will be able to quickly collect sets of "live" clinical data with the ability to filter by 150 parameters: gender, age, patient symptoms and other characteristics. This was announced on December 19, 2023 by representatives of Yandex Cloud.

Medical Data Platform in Numbers

According to the company, the platform is a web interface with search cells, in which you can search for documents and datasets by keywords, as well as apply various filters. For example, you can find all recorded cases of diabetes in people aged 20 to 50 years, collect examples of patients with elevated creatinine in a given range of values, unload all known orphan diseases, and much more. The system automatically selects relevant documents with links to X-rays, laboratory test results, CT, MRI and other tests. Data can be saved, uploaded in tabular form and used for AI development.

In 2024, more than 1000 specialists - university employees will receive access to the platform. The organization plans to gradually connect new data sources, customize data visualization using technology DataLens , and add the ability to work with natural language queries using generative. neuronet YandexGPT

Thanks to this platform, we will be able to collect data for scientific activities, educational work, and writing articles several orders of magnitude faster. What previously took months of painstaking work in the archives can now be done in less than a minute if you set the correct search parameters. We expect that in the future other medical organizations will be able to connect to the platform in order to exchange data and make scientific discoveries. We are also ready to share datacets with the scientific community and developers of artificial intelligence services, including on a commercial basis. comments Konstantin Brazhnikov, Deputy Director of the Department for the Development of Clinical and Educational Projects of the First Moscow State Medical University named after I.M. Sechenov Ministry of Health of the Russian Federation

This is a multi-format platform, with the ability to search in a single window for texts of medical conclusions, CT results, MRI, X-rays, laboratory test results and many other data necessary for the work of medical specialists. More than 18 million medical documents have been processed. These are large amounts of data that have been arranged and brought into a single format to allow full-text search. Another task was to organize constant data updates, which, of course, was helped by the cloud architecture of the solution. told Anna Lemyakina, Director of National Strategic Projects Yandex Cloud

From the very beginning, we saw the involvement of the Sechenov University team, felt the demand and potential utility of the analytical platform for the medical institution. Of course, this helped to complete the project in a fairly short time - nine stages, from the development of a web portal layout to the launch of the platform, were actually implemented in seven months. The team from our side included all the necessary specialists - backend and frontend developers, date engineers, cloudy architects, designer and technical writer, but the presence of a network and INFORMATION SECURITY expertise at the system integrator Beltel"," part of which is Beltel Datanomics. We are also grateful First Bit to the company for their assistance in designing the system for unloading primary data from the IIS. 1C narrated by Anna Plemyashova, director of Beltel Datanomics

The solution was developed and implemented by Beltel Datanomics, an expert in Big Data analytics and artificial intelligence. The university's platform uses more than 10 Yandex Cloud services: virtual machines are deployed using Compute Cloud. As of December 2023, there is not much data, the data warehouse is built on the Managed PostgreSQL managed service, as the load grows, migration to Managed Service for Greenplum is possible. The search system uses Managed OpenSearch technology. An important condition was to keep the "raw" data unchanged so that later they can be reprocessed, so the primary data and logs go to the S3 Object Storage. The information security system is built in accordance with the information security policy of the university.